What would you think of taking the audio file of some type of speech (a speech, an interview, a statement or anything) and editing it just by changing the words in the transcribed text? Sounds scary, doesn't it? For this is exactly the proposal of the Descript, an ultra-advanced audio editor for Mac from the same Groupon creator Andrew Mason.
In an interview with The VergeMason explained how his app works. According to the developer, the software generates a text description of the audio using speech recognition technologies; from then on, machine learning techniques come into play to automatically synchronize text with sound waves. Thus, each word is linked to a point in the timeline and, editing these words, the changes are reflected in the audio file when you delete a sentence from the text, for example, it will be automagically erased from speech.
Descript also includes a traditional sound wave editor for users to make changes to the manual form or adjust small imperfections in automatic editing.
Mason admits that his creation can be used for, say, unethical purposes after all, think about what it might mean for testimonies, evidence of crimes or other statements by important leaders and things like that. The creator of Descript, however, compares his creation to photo editors like Photoshop: ?Basically what happened for photos and prints is going to happen for audio and video, and society adjusts. The credibility of a piece of content is linked to the credibility of the source. ?
With all this technology built in, it was to be expected that Descript wasn't cheap and it really isn't. There are only two versions: the first costs US $ 20 / month (with an initial offer of US $ 10 monthly for the first months) and charges US $ 0.07 per minute of transcription; the second is $ 0.15 for the minute of transcription, with no monthly subscription. The software can be downloaded from its official website and brings 30 minutes of free transcripts for testing.
What do you think?